New Delhi, Apr 30: Genpact, an agentic and advanced technology solutions company, released a new research report with HFS Research, Autonomy Requires Trust in AI, that reveals a significant gap between agentic AI ambition and organizational readiness. The study explores the key barriers enterprises face when scaling agentic AI to uncover the foundational elements required to shift AI from a productivity tool to an autonomous execution layer within the enterprise: accountability, measurement, people, and process.
The research, based on a survey of 545 senior executives across 11 industries and interviews with leaders from Fortune 2000 companies, found that 92% of respondents believe agentic AI systems that can autonomously coordinate tasks and make decision will fundamentally change how work is executed. Despite this, nearly 80% of organizations still operate these systems in supervised modes, with humans retaining final approval over most actions.
“Enterprises have proven that generative AI can accelerate work; the next frontier is proving that agentic AI can take responsibility for executing it,” said Ajay Vasal, Global Leader for Data & AI, Genpact. “This shift from assistive to autonomous AI is beyond a technology upgrade and requires a fundamental rethink of operating models. Success depends on how organizations redesign processes with accountability and decision rights built in, while properly equipping their people to ensure agentic AI operates as a true extension of human intent.”
Key findings
The research points to four foundational elements determining whether agentic AI scales to autonomous execution or remains experimental:
- The trust and accountability gap: Only 22% of organizations are comfortable granting AI agents broad autonomy. The biggest barriers are not technical but center on regulatory exposure, reputational risk, lack of explainability, and unclear accountability.
- Investment rising, metrics lagging: Enterprises expect to scale agentic AI within an average of 17 months, with spending projected to increase by 38% in the coming year. However, 67% still rely on outdated productivity metrics that fail to capture the value of autonomous decision-making.
- Flatter organizations, shifting skills gap: 44% of executives expect fewer management layers as agentic AI absorbs coordination work. At the same time, the top skills in demand today are shifting from building AI to operating alongside it workflow orchestration and integration (42%), data engineering (39%), and monitoring and observability (36%) – meaning structural redesign for autonomy is moving faster than human role design.
- Process readiness is the leading barrier: 33% of respondents cite unprepared business processes as the top obstacle to agentic AI adoption, highlighting that workflow redesign – not technology – determines how far autonomy can scale.
Four critical actions for scaling autonomy
The report identifies four key actions to move beyond experimentation to enterprise-scale autonomous execution:
- Accountability: Defining who owns the agent’s actions and failures.
- Measurement: Adopting agent-native KPIs that track outcomes such as autonomous workflow completion.
- People: Providing role clarity and defining new human oversight responsibilities.
- Process: Redesigning workflows to remove manual handoffs and sequential human approvals.
“Ambition for autonomy is outpacing the ability to govern it,” said Phil Fersht, CEO and Chief Analyst, HFS Research. “The leaders in agentic AI won’t be the fastest adopters, they’ll be the ones who redesign accountability, measurement, and human oversight so autonomy scales with control and delivers compounding value.”
